Plutora Blog - Release Management
Managing Mixed Releases in Bimodal ITReading time 5 minutes
Bimodal IT is a trend identified by Gartner that describes a common challenge for IT departments – how to support two modes of operation within the same IT department. According to Gartner’s definition:
“Mode 1 is traditional and sequential, emphasizing safety and accuracy. Mode 2 is exploratory and nonlinear, emphasizing agility and speed.”
Bimodal IT is an accurate picture of the current reality of the IT department, but it’s often impossible to fully isolate both modes when it comes to releases. When you release software to production you are integrating systems across both modes. Release managers sit at the boundary between these two modes and are often responsible for managing scheduling and resource conflicts between Mode 1 and Mode 2 projects.
Here are just some of the ways Plutora is designed to support enterprise release management across Mode 1 and Mode 2 projects in a Bimodal IT department:
#1 – Plutora Release Manager models processes across Mode 1 and Mode 2
Our teams designed Plutora to be able to adapt to your process, and Plutora can support a wide range of release and deployment processes within the same enterprise release plan. We understand that today’s IT department spans a wide range of project management styles with agile (Mode 2) projects sitting alongside slower moving (Mode 1) projects. Often these two projects target the same testing environments and release timelines so understanding differences in capacity requirements, governance gates, and automation approaches is key to helping managers orchestrate across this “Bimodal boundary.”
When a release manager is planning a complex release process Plutora makes release dependencies easier to identify and Plutora’s system impact matrix will clearly show you the systems that are impacted by a release. Using this impact matrix you can identify what systems need to be prepared for a release, and in a Bimodal environment you can take steps to deal with dependencies spanning Mode 1 and Mode 2 systems. The key take-away is that Plutora doesn’t just assume that every project is agile, we give the release manager the tools they need to model variation across projects.
#2 – Plutora Environment Manager can model a wide range of environments
Real-world, enterprise systems span a wide range of technologies, and one of the biggest challenges facing the Bimodal IT department is how to synchronize environments between slower-moving Mode 1 projects and more agile Mode 2 projects. When a Mode 1 project requires two weeks of data preparation to begin the QA process for a quarterly release how is that requirement managed alongside projects that need to release every day.
Mode 1 projects may have long-lived, static environments dedicated to QA and Staging while Mode 2 projects may be embracing a fast-moving approach that involves on-demand cloud infrastructure. For example, your Mode 1 testing and integration database may be running on a static environment in a data center that hasn’t changed for years while your front-end teams may be making use of public clouds to create multiple environments on-demand that only exist for the duration of a single integration test run.
Plutora’s Environment Manager doesn’t ship with an assumption that all of your environments use the same, cookie-cutter approach. Plutora has the ability to manage both approaches within the same release timeline, and if your goal is to transition environments from Mode 1 to Mode 2 Plutora helps you identify opportunities for improvement.
#3 – Plutora models real capacity requirements across Mode 1 and Mode 2
Different projects require different resources. A Mode 1 system that conducts infrequent, quarterly releases likely requires an “all-hands-on-deck” approach to regression testing that can consume an entire QA department. At the opposite end of the spectrum, a Mode 2 application deployed to production multiple times a day may require only cursory tests run by an automated system. When dealing with a release timeline spanning Mode 1 and Mode 2 you need a tool that can assess the impact on limited testing and release engineering resources.
While system administrators and developers might think of environments and computing resources as the real capacity limits Enterprise Release Managers understand that the real capacity limits are often related to staff available to test, qualify, and execute a highly orchestrated release plan. Just as Plutora provides enterprise release managers with a way to model heterogeneous approaches to release processes Plutora also provides management with a window into real capacity. You can see how that behemoth Mode 1 project is going to take QA out of rotation for a few weeks, and you can prepare other projects to adapt to this reality.
Avoiding Bimodal Collisions
Most enterprises are running Bimodal IT departments whether they call it Bimodal or not, and resource allocation issues become apparent when a legacy system’s release schedule impacts a Mode 2 release schedule. These are the events that create unrealistic release schedules that impact both system availability and department morale. You can avoid these conflicts by using a tool that can understand and predict when a Mode 1 project is about to collide with a Mode 2 project.
We’ve talked about Plutora using Air Traffic Control analogies before and that analogy works for Bimodal. A release manager in a Bimodal environment is landing two types of projects (airplanes) on a limited number of environments (runway space.) Your challenge is to manage high-risk Mode 1 projects requiring the same resources as faster moving agile projects all impacting the same teams and systems to achieve a release that won’t impact system availability.
This high-wire act is analogous to dealing with several incoming Boeing 747s or Airbus 380s on the same runway space that is supporting smaller jets like a Gulfstream G650. These planes can’t all land on the runway at the same time and they have different requirements for gates and support. In a Bimodal environment, you are dealing with a similar challenge. How do you model the release pipeline and assess system impact so that you can make adjustments to Mode 1 and Mode 2 projects while they are in transit. (Alternatively, how do you predict and prevent conflicts before that Gulfstream is only 1000 feet away from that 747?)